There are many applications on computing devices which require selection from a list where the item identifier for the list is text based. These include selecting a contact from an address book, selecting a record from a database, selecting a word from a dictionary, etc.
The user of such systems has to enter enough information to narrow a search of the list down to a single entry to be selected. This usually entails entering several of the initial letters of the search term resulting in a smaller subset of the list, then scrolling through the subset to manually point out the desired item.
This can be a cumbersome process requiring the user to be able to enter any letter in their language's alphabet as well as possibly numbers and other symbols. This is particularly onerous on small, hand held devices such as mobile phones, personal digital assistants (PDAs), remote controls, etc.
On hand held devices with a keypad such as mobile phones (FIG. 1), the try of alphabetic characters as well as numerics and symbols is achieved through a cumbersome system of multiply pressing the 12 number keys (0 to 9 and “*” and “#”). This method has the disadvantages of:                Being a complicated, unnatural system which needs to be learnt,        Requiring accurate presses of generally very small buttons resulting in accidental pressing of adjacent buttons,        Requiring the user to be able to read and discern very small labels on the buttons,        Requiring constant shift of gaze between buttons and screen to track the input,        Only allowing the entry of one character at a time, not a group of characters at a time.        
Other handheld devices take input through a touch screen such as PDAs (FIG. 2). These allow text input through generally two methods: an on-screen keyboard and handwriting recognition.
The on-screen keyboard method involves presenting an image of a keyboard on the screen. The buttons on this on-screen keyboard are too small to be selected using fingers so a stylus is required. This results in a slow, error prone selection of individual characters with the disadvantages:                Access to numbers and additional symbols usually requiring at least two presses as they require the keyboard to go into another mode to allow access to them,        Small key images are hard to see for those with any impairment of vision and hard to select for those with limited dexterity,        Characters are only entered one at a time.        
Handwriting recognition systems have the disadvantage of a high error rate in recognising characters unless either the system is trained with the user's handwriting or the user learns a set of distinctive ways to input letters that are amenable to the device's programming. This method has the disadvantages of:                Needing a stylus for input,        Being slow to enter each letter,        Requiring a constant cycle of entry, checking of accuracy of the entry and correction of misinterpretation of entries,        Characters are only entered one at a time.        
Even using a fully fledged keyboard to specify the characters to select an object has disadvantages:                It is still necessary to enter one character at a time,        It is still necessary to enter enough characters to generate a manageable subset list and then select from it, the alternative would be to enter sufficient characters each time to uniquely identify the entry which may require almost all of the entry to be specified,        Except for adept touch typists there is still a need to constantly move one's gaze between the keyboard and the screen.        
There is need for an improved method of selecting items from a list.